Medicolegal Issues

In the Literature


Repeat Testing for C. Diff?

By Jeff Glasheen, MD

Mohan SS, McDermott BP, Parchuri S, et al. Lack of value of repeat stool testing for Clostridium difficile toxin. Am J Med. 2006 Apr;119(4):356.e7-356.e8

Clostridium difficile is a common complication of antibiotic and chemotherapeutic use, especially in hospitalized patients. Yet most nosocomial diarrhea is not caused by C. difficile. Most antibiotics can cause loose stools through changes in the gastrointestinal flora that result in inadequate digestion and absorption of carbohydrates and a resultant osmotic diarrhea. Further, antibiotics such as erythromycin and amoxicillin/clavulanate may result in diarrhea via increases in GI tract motility. While osmotic and motility causes of diarrhea tend to improve with antibiotic discontinuation, C. difficile-associated diarrhea is associated with significant morbidity that often continues until adequately treated.

This study’s authors conclude that repeat “C. difficile testing is not clinically justified and is economically wasteful.”

Thus having a test that differentiates between C. difficile and non-clostridial diarrhea is essential. The most commonly used test is the enzyme immunoassay (EIA) that detects toxins A and B. The sensitivity and specificity of this test has been reported to range between 50%-90% and 70%-95%, respectively. The authors of this paper evaluated the utility of repeat EIA testing in patients with a one negative test in the setting of nosocomial diarrhea associated with antibiotic use.

The authors reviewed 474 sequential EIA tests for C. difficile in 396 patients over a 10-month period at a large university-affiliated community hospital with an EIA sensitivity and specificity of 80%-90% and 80%-95%, respectively. Tests were considered to be “repeat” if they occurred within seven days of the original negative test. Of the 78 repeat tests (16.5% of all tests), only one was positive, resulting in a 0.8% conversion rate. At an institutional cost of $128 per test the total cost of EIA testing over the 10-month period was $60,672. The cost of repeat testing alone was $9,984. The authors conclude that there is limited value—and high cost—in repeat EIA testing and that alternative sources of diarrhea should be sought or we should repeat EIA testing in patients with continued nosocomial diarrhea and a negative EIA test.

While prior studies have shown incremental benefit of retesting for C. difficile with the EIA assay, this study’s authors conclude that repeat “C. difficile testing is not clinically justified and is economically wasteful.” Unfortunately, the authors did not utilize a strong enough research design to defend this statement. From the data presented, all we can conclude is that repeat testing with an EIA assay did not add significant value to the diagnostic workup. The lack of a negative gold standard test, such as the cytotoxin assay or follow up outcomes, such as resolution of diarrhea, leaves the reader pondering if the repeat EIA assays were negative because the patients did not have C. difficile or because the test was not sensitive enough to detect the toxin. The reported sensitivity for the EIA assay used was 80-90%, meaning that 10%-20% of patients with C. difficile disease would have had a false negative test. While a second negative test would lower the likelihood of true disease, it would not lower it enough in a patient with a high pre-test probability of disease to sufficiently rule out the disease such that further testing is indicated.

Additionally, while we can extrapolate that the EIA test was of little utility to the patients studied here, no patient-specific data is presented. Thus it is difficult to determine if our patient population is represented in the study. More knowledge about the patients would allow the reader to use published prediction rules to better delineate how likely it was that this cohort was at high risk for having toxin-mediated diarrhea. Perhaps a more reasonable approach to this clinical conundrum would be to send the repeat stool test for a cytotoxin assay or to treat the patient empirically in situations where either the likelihood of disease or the disease burden is high.

The review included studies spanning more than 25 years, and thus included data from studies done in the era prior to modern therapy for cardiac ischemia including potent antiplatelet and statin therapy.

Stroke after MI

By Ethan Cumbler, MD

Witt BJ, Ballman KV, Brown RD Jr, et al. The incidence of stroke after myocardial infarction: a meta-analysis. Am J Med. 2006 Apr;119(4):354.e1-9.

Stroke and myocardial infarction (MI) share many of the risk factors leading to atherosclerosis, including hypertension, hyperlipidemia, diabetes, tobacco abuse, and age. Logically, patients at risk for one event would also be at risk for the other, yet this interaction appears to be more complex. The study by Witt, et al., aims to elucidate the rate of in-hospital stroke in patients initially admitted with an MI.

The authors analyzed 22 observational studies of myocardial infarction that recorded the incidence of cerebrovascular accidents after acute MI. Clinical trials were excluded from the analysis in order to provide representation of an unselected population. Of the trials reviewed, 11 were included in the analysis of in-hospital strokes, three for the 30-day time point, and two for the one-year time point. The other trials used different time points and were not included in the analysis. The patients had a mean age ranging between 59 and 72.7 years, and all had a predominance of males. The rate of in-hospital stroke was 11.1 events per 1,000 hospitalizations. This incidence rose to 12.2/1,000 at 30 days and 21.4/1,000 at one year.

Plausible hypotheses for why the rate of cerebrovascular accident would be particularly high in the post-MI period include the potential for localized wall motion dysmotility or low flow leading to intracardiac thrombosis, event-related arrhythmia, or procedure-related embolic events. The studies from which this meta-analysis was derived were not designed in such a way for a causal relationship to be identified. However, age, diabetes, hypertension, prior stroke, anterior location of MI, atrial fibrillation, heart failure, and nonwhite race were all found to have an association with increased risk for stroke. Interestingly, angina on presentation was associated with an apparent decreased risk, theorized to potentially be due to ischemic preconditioning.

While this study shares the usual limitations of meta-analyses of observational studies, the authors have done an excellent summation of the data available including both English and non-English language articles in the analysis. Notably, the review included studies spanning more than 25 years and, thus, included data from studies done in the era prior to modern therapy for cardiac ischemia including potent antiplatelet and statin therapy. The three studies with in-hospital time points started in the 1990s had a lower average rate of stroke, which may reflect the effect of more potent anti-platelet agents used in today’s therapy for acute coronary syndromes.

The implication for the hospitalist is to recognize that patients admitted for MI are at high risk for stroke during the index hospitalization. A low threshold for suspicion of a cerebrovascular event needs to be maintained for post-MI patients with new neurologic symptoms. Future studies will be needed to address the risk/benefit of anticoagulation in high-risk patients for stroke following a myocardial infarction.

VARIABLE / POINTS click for large version

click for large version

Predicting PE in the ED Using Revised Geneva Score

By Jeffrey Carter, MD

Le Gal G, Righini M, Roy PM, et al. Prediction of pulmonary embolism in the emergency department: the revised Geneva score. Ann Intern Med. 2006 Feb 7;144(3):165-171.

Introduction: Pulmonary embolism remains a common life-threatening disorder with imperfect diagnostic modalities and strategies. Much of the current literature focuses on the development and validation of clinical probability assessments that identify low-risk patients who can be safely managed without invasive testing or lung scanning.

Two common scoring systems include the Geneva and Wells criteria, which employ a combination of historical information, clinical and laboratory data to stratify patients into three risk groups.1,2 It has recently been shown that a single negative D-dimer test safely excludes patients with suspected VTE in a large cohort, with a three-month follow-up rate of VTE of 0%.3 Criticism of current clinical probability assessment strategies include the inclusion of subjective criteria or the need for blood gas values.1,2 In this paper, the authors sought to derive and validate a scoring system based on readily available objective clinical data.

Methods: The cohort evaluated in this study consists of the same 965 patients used to prospectively evaluate an emergency department diagnostic strategy of VTE.3 Patients presenting to the emergency ward with complaints of chest pain or shortness of breath were evaluated with a standard diagnostic workup that included a checklist of demographic and historical information, signs and symptoms possibly consistent with VTE, ABG, chest X-ray, and EKG results, as well as the likelihood of another diagnosis. However, the goal was to create a scoring system based on readily available objective data, so ABG, chest X-ray and electrocardiogram results were not evaluated as possible components of the score. Clinical variables associated with pulmonary embolism were evaluated for statistical significance in both a univariate and multivariate model. These variables were then validated both internally and with an external cohort.

Results: The incidence of VTE was 23% (222/965). Ten clinical variables were found to be statistically significantly associated with VTE; these comprise the Revised Geneva Score.

Clinical probability is based on points: 0-3 is low probability; 4-10 is intermediate probability, and >10 is high probability. Low, intermediate, and high clinical probabilities had respective rates of VTE of 8%, 28%, and 74%. These percentages were found in the validation cohort and were similar to those in the derivation cohort.

Discussion: This study focuses on the derivation and validation of a clinical scoring system that can provide a numerical estimate of likelihood of VTE. The diagnostic strategy by which VTE is confirmed or excluded is from derivation cohort and is published elsewhere.3 In that study patients classified as no VTE and not treated had a 90-day risk of VTE of 1%. The nine-point revised Geneva score does indeed provide an accurate assessment of risk of VTE, and can thus help guide clinical decision-making. However it is not clear that the revised Geneva score will help decrease invasive diagnostic tests. In the low-risk group, 8% is too great a risk of a life-threatening illness to forego further diagnostic evaluation. Coupled with a negative ELISA D-dimer, exclusion of these patients is safe, but these patients cannot be safely excluded without risk stratification.3


  1. Wells PS, Anderson DR, Rodger M, et al. Derivation of a simple clinical model to categorize patients’ probability of pulmonary embolism: increasing the model’s utility with the SimpliRED D-dimer. Thromb Haemost. 2000 Mar;83:416-420.
  2. Wicki J, Perneger TV, Junof AF, et al. Assessing clinical probability of pulmonary embolism in the emergency ward: a simple score. Arch Intern Med. 2001 Jan 8;161(1):92-97.
  3. Perrier A, Roy PM, Aujesky D, et al. Diagnosing pulmonary embolism in outpatients with clinical assessment, D-dimer measurement, venous ultrasound and helical computed tomography: a multicenter management study. Am J Med. 2004 Mar 1;116(5):291-299.
This study focuses on the derivation and validation of a clinical scoring system that can provide a numerical estimate of the likelihood of VTE.

LMWH for Inpatient Palliative Care

By Jeanie Youngwerth, MD

Noble SI, Nelson A, Turner C, et al. Acceptability of low molecular weight heparin thromboprophylaxis for patients receiving palliative care: qualitative study. BMJ. 2006 Mar 11:332(7541);577-580.ePub 2006 Feb 3.

Venous thromboembolism is a major risk factor for patients with malignancy. VTE may reduce survival time in patients receiving palliative care, with one in seven inpatients with cancer dying from pulmonary embolism. The American College of Chest Physicians recommend low molecular weight heparin (LMWH) in hospitalized patients with cancer as level 1A evidence in their 2004 consensus statement on VTE. There are no thromboprophylaxis guidelines in the United Kingdom. Many physicians view daily injections of LMWH as unnecessary distress for palliative care patients. This study focused on what inpatients with advanced cancer receiving palliative care thought about the effect of thromboprophylaxis on overall quality of life.

This was a qualitative study of 28 inpatients receiving palliative care who had advanced metastatic cancer and who were in a regional cancer center in Wales. The patients had received LMWH for at least five consecutive days. The patients were audiotaped and then had semi-structured interviews transcribed regarding cancer treatments they had received. These interviews covered the patients’ insight into prognosis, their understanding of thromboprophylaxis, their beliefs concerning the effects of thromboprophylaxis on overall quality of life, and the negative aspects of heparin treatment.

The main outcome measures were recurring themes of the effect of thromboprophylaxis on overall quality of life. Major themes identified were insight into prognosis, knowledge and understanding of thromboprophylaxis, acceptability of heparin thromboprophylaxis, reassurance, and optimism. Minor themes identified were bruising, negative impact of antiembolic stockings, and anger at paternalistic views toward terminally ill patients.

Most patients showed clear insight into the nature of their condition and understood heparin prophylaxis for VTE. Many patients identified immobility and surgery as VTE risk factors, with little understanding of cancer as a risk factor. All knew that VTE could cause death, but none were aware of the common symptoms. All patients found LMWH thromboprophylaxis acceptable. Patients believed taking measures to prevent symptoms might improve their quality of life and felt that the medical team had not given up on them.

The only negative experiences of LMWH thromboprophylaxis was bruising. All of the patients who wore antiembolic stockings during previous admissions found them uncomfortable. Patients expressed their need to be involved in decision-making, particularly with respect to withdrawal or non-administration of treatment.

This small, qualitative study showed that, contrary to many physicians’ beliefs, patients in palliative care units believe that LMWH injections are acceptable as thromboprophylaxis, but antiembolic stockings are not. The effect of daily injections had little or no effect on the quality of life, with many patients having positive feelings that things were being done to prevent new problems from occurring.

This study was limited by the small sample size, the qualitative nature that could introduce interpretation bias, and the fact that only patients using LMWH were included. Additionally, the United Kingdom has aggressively educated the public on the risks of VTE associated with long flights such that baseline knowledge may differ in other parts of the world. Resource utilization, including drug costs and length of stay, as well as effect on mortality were not studied.

The implications for hospitalists are that many inpatients with advanced cancer receiving palliative care may find LMWH thromboprophylaxis acceptable therapy, and that discussions with the patient regarding this option should be explored together.

Intensive control of hyperglycemia in the hospitalized ICU patient appears to be associated with reduction of morbidity and mortality.

Cost Analysis: Intensive Glycemic Control in Critically Ill Adults

By Whitney Woodmansee, MD

Krinsley JS, Jones, RL. Cost analysis of intensive glycemic control in critically ill adult patients. Chest. 2006;129:644-650.

Several studies have demonstrated improved outcomes and decreased mortality in ICU patients treated with intensive control of blood glucose levels. This study sought to identify the costs associated with intensive glycemic control in the ICU.

An ICU patient database was analyzed for cost data related to intensive glycemic control. A baseline group of 800 consecutive ICU patients admitted prior to initiation (baseline) of an intensive glucose management (blood glucose levels between 80-140 mg/dl protocol were compared with a treatment group of 800 consecutive patients admitted after initiation of the protocol). Previously reported outcomes of these patients demonstrated significant improvement in mortality with tight glycemic control. Costs related to ICU and non-ICU length of stay; duration of mechanical ventilation; and all laboratory, pharmacy, and radiology services were analyzed between groups. Resource utilization was determined by assessing charges from the database and adjusting for inflation and applying Medicare cost, charge ratios for each category, and fiscal year. Costs associated with the intensive glycemic control protocol were determined. Unfortunately, only cost estimates for insulin and associated disposable supplies for each group were available for analysis.

Baseline and treatment populations did not differ significantly regarding demographics such as age, gender, race, admitting diagnosis, diabetes prevalence, or APACHE II scores. There were fewer patients in the treatment group that required mechanical ventilation during their ICU stay (40.6% versus 33.6%). Intensive glucose management was associated with a 13.9% reduction in total ICU days and duration of mechanical ventilation (median of two days decreased to 1.7 days p=0.045). There was a $1,580 adjusted cost savings per patient in the intensive treatment group compared with the baseline group (p<0.001). This reduction in cost was primarily driven by a decrease in laboratory and radiology costs in the ventilated patients. There were nonsignificant cost reductions in the unventilated patients.

Intensive control of hyperglycemia in the hospitalized ICU patient appears to be associated with reduction of morbidity and mortality. This suggests that tight glycemic control also leads to reductions in overall patient care costs—particularly in the ventilated ICU patient. Although not a randomized control trial, database analysis of costs and resource utilization demonstrated an overall cost savings in the treatment group (after initiation of an intensive glycemic control protocol) compared with the baseline group (before protocol initiation). One caveat is that the authors used estimates when determining the costs associated with the implementation of the intensive glucose management protocol. Nevertheless, intensive glycemic control was associated with an overall reduction in patient costs related to decreased ICU days and mechanical ventilation as well as resource utilization in a patient population already shown to have improved mortality. These results, if confirmed, suggest that tight glycemic control in the ICU is cost effective and should become standard medical practice. TH

Classic Literature

Beat the Heroin Habit

By William Rifkin, MD

Dole VP, Nyswander M. A medical treatment of diacetylmorphine (heroin) addiction. JAMA. 1965 Aug 23;193:80-84.

This article is the 12th most cited article from JAMA in the 1960s. It had been cited 473 times as of April 2006, an average of almost once a month since publication. Forty-seven citations have occurred since 2000, most recently in January 2006.


The authors examined “whether a narcotic medicine, prescribed as part of a treatment program, could help in the return of addict patients to normal society.” This was the first study on such a practice.

The patients were simply described as “22 non-psychotic heroin users … with a history of failures of withdrawal treatment.” They were further classified as Western European, Puerto Rican/Cuban, or black.

Examined range of time in methadone treatment was one week to 15 months.


Presented in narrative format; “The most dramatic effect … has been the disappearance of narcotic hunger.” Further, “ … patients found that they could meet addict friends and even watch them inject without great difficulty. … They have stopped dreaming about drugs … have even become so indifferent to narcotics as to forget to take a scheduled dose … .”

The degree of tolerance was tested in six patients, “ … by giving heroin, morphine, Dilaudid or methadone intravenously in a double blind study.”

They described the use of non-prescribed narcotics as “ … unscheduled, but perhaps necessary, experiments in drug usage made by four patients. These subjects found that they did not ‘get high’ … patients and their friends were astounded at their lack of reaction … .”

As an aside the authors note that constipation was assessed by giving five patients a barium sulfate meal and following daily X-ray exams for a week.


“Maintenance with methadone is no more difficult than maintaining diabetics with oral hypoglycemic agents … .”

One should note that this was an uncontrolled case series. As such it is very low in the hierarchy of study strength to make firm conclusions. Yet the authors do, most emphatically by today’s standards.

I am astounded by the difference of acceptable research behaviors and standards 40 years ago with those of today. Firstly, what were the ethics of giving patients IV heroin, especially with no mention of informed consent or an IRB. Secondly, note the cavalier attitudes to radiation evidenced by a side examination of constipation via a barium meal and daily X-rays.

Finally, I am quite surprised that a study with the below outlined threats to validity, which, although it could be argued render the data meaningless, is one of the most cited and referenced articles in the whole medical literature, let alone in the specific area of narcotic replacement therapy.

By today’s standards for scientific rigor, one should note that no data were presented. This was simply a descriptive narrative of experiences as reported by the physician authors of their patients. Further the median duration in the program was only three months. Ten of the 22 patients were treated for less than six weeks—two for less than 14 days. Nevertheless, the authors make dramatic conclusions about a chronic disease.

Obviously we should be careful about applying 21st century standards of science and ethics to a 40-year-old study. However, it does give pause that at least as recently as 1965 it appears researchers were granted relatively free rein by society to perform studies on patients and by editors of a medical publication to make dramatic conclusions.

In the case of methadone, of course, time has tempered the exuberant optimism on display here. Narcotic replacement therapy is clearly useful, but, alas, patients hardly “… forget about scheduled doses …” and treatment would not be classified as “… no more difficult than maintaining diabetics with oral hypoglycemic agents … .” TH

Next Article:

   Comments ()